from transformers.utils import ModelOutput import torch from torch import nn from typing import Dict, List, Tuple, Optional from dataclasses import dataclass from transformers import BertPreTrainedModel, BertModel, BertTokenizerFast # define the classes, and the possible prefixes for each class POSSIBLE_PREFIX_CLASSES = [ ['לכש', 'כש', 'מש', 'בש', 'לש'], ['מ'], ['ש'], ['ה'], ['ו'], ['כ'], ['ל'], ['ב'] ] POSSIBLE_RABBINIC_PREFIX_CLASSES = [ ['לכש', 'כש', 'מש', 'בש', 'לש', 'לד', 'בד', 'מד', 'כד', 'לכד'], ['מ'], ['ש', 'ד'], ['ה'], ['ו'], ['כ'], ['ל'], ['ב'], ['א'], ['ק'] ] class PrefixConfig(dict): def __init__(self, possible_classes, **kwargs): # added kwargs for previous version where all features were kept as dict values super().__init__() self.possible_classes = possible_classes self.total_classes = len(possible_classes) self.prefix_c2i = {w: i for i, l in enumerate(possible_classes) for w in l} self.all_prefix_items = list(sorted(self.prefix_c2i.keys(), key=len, reverse=True)) @property def possible_classes(self) -> List[List[str]]: return self.get('possible_classes') @possible_classes.setter def possible_classes(self, value: List[List[str]]): self['possible_classes'] = value DEFAULT_PREFIX_CONFIG = PrefixConfig(POSSIBLE_PREFIX_CLASSES) def get_prefixes_from_str(s, cfg: PrefixConfig, greedy=False): # keep trimming prefixes from the string while len(s) > 0 and s[0] in cfg.prefix_c2i: # find the longest string to trim next_pre = next((pre for pre in cfg.all_prefix_items if s.startswith(pre)), None) if next_pre is None: return yield next_pre # if the chosen prefix is more than one letter, there is always an option that the # prefix is actually just the first letter of the prefix - so offer that up as a valid prefix # as well. We will still jump to the length of the longer one, since if the next two/three # letters are a prefix, they have to be the longest one if not greedy and len(next_pre) > 1: yield next_pre[0] s = s[len(next_pre):] def get_prefix_classes_from_str(s, cfg: PrefixConfig, greedy=False): for pre in get_prefixes_from_str(s, cfg, greedy): yield cfg.prefix_c2i[pre] @dataclass class PrefixesClassifiersOutput(ModelOutput): loss: Optional[torch.FloatTensor] = None logits: Optional[torch.FloatTensor] = None hidden_states: Optional[Tuple[torch.FloatTensor]] = None attentions: Optional[Tuple[torch.FloatTensor]] = None class BertPrefixMarkingHead(nn.Module): def __init__(self, config) -> None: super().__init__() self.config = config if not hasattr(config, 'prefix_cfg') or config.prefix_cfg is None: setattr(config, 'prefix_cfg', DEFAULT_PREFIX_CONFIG) if isinstance(config.prefix_cfg, dict): config.prefix_cfg = PrefixConfig(config.prefix_cfg['possible_classes']) # an embedding table containing an embedding for each prefix class + 1 for NONE # we will concatenate either the embedding/NONE for each class - and we want the concatenate # size to be the hidden_size prefix_class_embed = config.hidden_size // config.prefix_cfg.total_classes self.prefix_class_embeddings = nn.Embedding(config.prefix_cfg.total_classes + 1, prefix_class_embed) # one layer for transformation, apply an activation, then another N classifiers for each prefix class self.transform = nn.Linear(config.hidden_size + prefix_class_embed * config.prefix_cfg.total_classes, config.hidden_size) self.activation = nn.Tanh() self.classifiers = nn.ModuleList([nn.Linear(config.hidden_size, 2) for _ in range(config.prefix_cfg.total_classes)]) def forward( self, hidden_states: torch.Tensor, prefix_class_id_options: torch.Tensor, labels: Optional[torch.Tensor] = None) -> Tuple[torch.FloatTensor, torch.FloatTensor]: # encode the prefix_class_id_options # If input_ids is batch x seq_len # Then sequence_output is batch x seq_len x hidden_dim # So prefix_class_id_options is batch x seq_len x total_classes # Looking up the embeddings should give us batch x seq_len x total_classes x hidden_dim / N possible_class_embed = self.prefix_class_embeddings(prefix_class_id_options) # then flatten the final dimension - now we have batch x seq_len x hidden_dim_2 possible_class_embed = possible_class_embed.reshape(possible_class_embed.shape[:-2] + (-1,)) # concatenate the new class embed into the sequence output before the transform pre_transform_output = torch.cat((hidden_states, possible_class_embed), dim=-1) # batch x seq_len x (hidden_dim + hidden_dim_2) pre_logits_output = self.activation(self.transform(pre_transform_output))# batch x seq_len x hidden_dim # run each of the classifiers on the transformed output logits = torch.cat([cls(pre_logits_output).unsqueeze(-2) for cls in self.classifiers], dim=-2) loss = None if labels is not None: loss_fct = nn.CrossEntropyLoss() loss = loss_fct(logits.view(-1, 2), labels.view(-1)) return (loss, logits) class BertForPrefixMarking(BertPreTrainedModel): def __init__(self, config): super().__init__(config) self.bert = BertModel(config, add_pooling_layer=False) self.dropout = nn.Dropout(config.hidden_dropout_prob) self.prefix = BertPrefixMarkingHead(config) # Initialize weights and apply final processing self.post_init() def forward( self, input_ids: Optional[torch.Tensor] = None, attention_mask: Optional[torch.Tensor] = None, token_type_ids: Optional[torch.Tensor] = None, prefix_class_id_options: Optional[torch.Tensor] = None, position_ids: Optional[torch.Tensor] = None, labels: Optional[torch.Tensor] = None, head_mask: Optional[torch.Tensor] = None, inputs_embeds: Optional[torch.Tensor] = None, output_attentions: Optional[bool] = None, output_hidden_states: Optional[bool] = None, return_dict: Optional[bool] = None, ): r""" labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*): Labels for computing the token classification loss. Indices should be in `[0, ..., config.num_labels - 1]`. """ return_dict = return_dict if return_dict is not None else self.config.use_return_dict bert_outputs = self.bert( input_ids, attention_mask=attention_mask, token_type_ids=token_type_ids, position_ids=position_ids, head_mask=head_mask, inputs_embeds=inputs_embeds, output_attentions=output_attentions, output_hidden_states=output_hidden_states, return_dict=return_dict, ) hidden_states = bert_outputs[0] hidden_states = self.dropout(hidden_states) loss, logits = self.prefix.forward(hidden_states, prefix_class_id_options, labels) if not return_dict: return (loss,logits,) + bert_outputs[2:] return PrefixesClassifiersOutput( loss=loss, logits=logits, hidden_states=bert_outputs.hidden_states, attentions=bert_outputs.attentions, ) def predict(self, sentences: List[str], tokenizer: BertTokenizerFast, padding='longest'): # step 1: encode the sentences through using the tokenizer, and get the input tensors + prefix id tensors inputs = encode_sentences_for_bert_for_prefix_marking(tokenizer, self.config.prefix_cfg, sentences, padding) inputs.pop('offset_mapping') inputs = {k:v.to(self.device) for k,v in inputs.items()} # run through bert logits = self.forward(**inputs, return_dict=True).logits return parse_logits(inputs['input_ids'].tolist(), sentences, tokenizer, logits, self.config.prefix_cfg) def parse_logits(input_ids: List[List[int]], sentences: List[str], tokenizer: BertTokenizerFast, logits: torch.FloatTensor, config: PrefixConfig): # extract the predictions by argmaxing the final dimension (batch x sequence x prefixes x prediction) logit_preds = torch.argmax(logits, axis=3).tolist() ret = [] for sent_idx,sent_ids in enumerate(input_ids): tokens = tokenizer.convert_ids_to_tokens(sent_ids) ret.append([]) for tok_idx,token in enumerate(tokens): # If we've reached the pad token, then we are at the end if token == tokenizer.pad_token: continue if token.startswith('##'): continue # combine the next tokens in? only if it's a breakup next_tok_idx = tok_idx + 1 while next_tok_idx < len(tokens) and tokens[next_tok_idx].startswith('##'): token += tokens[next_tok_idx][2:] next_tok_idx += 1 prefix_len = get_predicted_prefix_len_from_logits(token, logit_preds[sent_idx][tok_idx], config) if not prefix_len: ret[-1].append([token]) else: ret[-1].append([token[:prefix_len], token[prefix_len:]]) return ret def encode_sentences_for_bert_for_prefix_marking(tokenizer: BertTokenizerFast, config: PrefixConfig, sentences: List[str], padding='longest', truncation=True): inputs = tokenizer(sentences, padding=padding, truncation=truncation, return_offsets_mapping=True, return_tensors='pt') # create our prefix_id_options array which will be like the input ids shape but with an addtional # dimension containing for each prefix whether it can be for that word prefix_id_options = torch.full(inputs['input_ids'].shape + (config.total_classes,), config.total_classes, dtype=torch.long) # go through each token, and fill in the vector accordingly for sent_idx, sent_ids in enumerate(inputs['input_ids']): tokens = tokenizer.convert_ids_to_tokens(sent_ids) for tok_idx, token in enumerate(tokens): # if the first letter isn't a valid prefix letter, nothing to talk about if len(token) < 2 or not token[0] in config.prefix_c2i: continue # combine the next tokens in? only if it's a breakup next_tok_idx = tok_idx + 1 while next_tok_idx < len(tokens) and tokens[next_tok_idx].startswith('##'): token += tokens[next_tok_idx][2:] next_tok_idx += 1 # find all the possible prefixes - and mark them as 0 (and in the possible mark it as it's value for embed lookup) for pre_class in get_prefix_classes_from_str(token, config): prefix_id_options[sent_idx, tok_idx, pre_class] = pre_class inputs['prefix_class_id_options'] = prefix_id_options return inputs def get_predicted_prefix_len_from_logits(token, token_logits, config: PrefixConfig): # Go through each possible prefix, and check if the prefix is yes - and if # so increase the counter of the matched length, otherwise break out. That will solve cases # of predicting prefix combinations that don't exist on the word. # For example, if we have the word ושכשהלכתי and the model predict ו & כש, then we will only # take the vuv because in order to get the כש we need the ש as well. # Two extra items: # 1] Don't allow the same prefix multiple times # 2] Always check that the word starts with that prefix - otherwise it's bad # (except for the case of multi-letter prefix, where we force the next to be last) cur_len, skip_next, last_check, seen_prefixes = 0, False, False, set() for prefix in get_prefixes_from_str(token, config): # Are we skipping this prefix? This will be the case where we matched כש, don't allow ש if skip_next: skip_next = False continue # check for duplicate prefixes, we don't allow two of the same prefix # if it predicted two of the same, then we will break out if prefix in seen_prefixes: break seen_prefixes.add(prefix) # check if we predicted this prefix if token_logits[config.prefix_c2i[prefix]]: cur_len += len(prefix) if last_check: break skip_next = len(prefix) > 1 # Otherwise, we predicted no. If we didn't, then this is the end of the prefix # and time to break out. *Except* if it's a multi letter prefix, then we allow # just the next letter - e.g., if כש doesn't match, then we allow כ, but then we know # the word continues with a ש, and if it's not כש, then it's not כ-ש- (invalid) elif len(prefix) > 1: last_check = True else: break return cur_len